Evaluation of Groundwater Storage Variations Estimated from GRACE Data Assimilation and State-of-the-Art Land Surface Models in Australia and the North China Plain

Remote Sensing

Edwin H. Sutanudjaja¹  Hannes Müller Schmied²  Siyuan Tian³  Natthachet Tangdamrongsub⁴  Shin-Chan Han⁴  Wei Feng⁵  Jiangjun Ran⁵ 

[1] Department of Physical Geography, Faculty of Geosciences, Utrecht University, 3584 CS Utrecht, The Netherlands;Institute of Physical Geography, Goethe-University Frankfurt, 60438 Frankfurt , Germany;Research School of Earth Sciences, Australian National University, Canberra, ACT 2601, Australia;School of Engineering, University of Newcastle, Callaghan, NSW 2308, Australia;State Key Laboratory of Geodesy and Earth’s Dynamics, Institute of Geodesy and Geophysics, Chinese Academy of Sciences, Wuhan 430077, China;

关键词: groundwater storage;    Australia;    North China Plain;    GRACE;    data assimilation;    EnKS 3D;    PCR-GLOBWB;    CABLE;    WGHM;    W3;   

DOI  :  10.3390/rs10030483

来源: DOAJ

【 摘 要 】

The accurate knowledge of the groundwater storage variation (ΔGWS) is essential for reliable water resource assessment, particularly in arid and semi-arid environments (e.g., Australia, the North China Plain (NCP)) where water storage is significantly affected by human activities and spatiotemporal climate variations. The large-scale ΔGWS can be simulated from a land surface model (LSM), but the high model uncertainty is a major drawback that reduces the reliability of the estimates. The evaluation of the model estimate is then very important to assess its accuracy. To improve the model performance, the terrestrial water storage variation derived from the Gravity Recovery And Climate Experiment (GRACE) satellite mission is commonly assimilated into LSMs to enhance the accuracy of the ΔGWS estimate. This study assimilates GRACE data into the PCRaster Global Water Balance (PCR-GLOBWB) model. The GRACE data assimilation (DA) is developed based on the three-dimensional ensemble Kalman smoother (EnKS 3D), which considers the statistical correlation of all extents (spatial, temporal, vertical) in the DA process. The ΔGWS estimates from GRACE DA and four LSM simulations (PCR-GLOBWB, the Community Atmosphere Biosphere Land Exchange (CABLE), the Water Global Assessment and Prognosis Global Hydrology Model (WGHM), and World-Wide Water (W3)) are validated against the in situ groundwater data. The evaluation is conducted in terms of temporal correlation, seasonality, long-term trend, and detection of groundwater depletion. The GRACE DA estimate shows a significant improvement in all measures, notably the correlation coefficients (respect to the in situ data) are always higher than the values obtained from model simulations alone (e.g., ~0.15 greater in Australia, and ~0.1 greater in the NCP). GRACE DA also improves the estimation of groundwater depletion that the models cannot accurately capture due to the incorrect information of the groundwater demand (in, e.g., PCR-GLOBWB, WGHM) or the unavailability of a groundwater consumption routine (in, e.g., CABLE, W3). In addition, this study conducts the inter-comparison between four model simulations and reveals that PCR-GLOBWB and CABLE provide a more accurate ΔGWS estimate in Australia (subject to the calibrated parameter) while PCR-GLOBWB and WGHM are more accurate in the NCP (subject to the inclusion of anthropogenic factors). The analysis can be used to declare the status of the ΔGWS estimate, as well as itemize the possible improvements of the future model development.

【 授权许可】

Unknown